The whole-house idea is overambitious, but, on a smaller scale, it would be a great way to recharge portable devices. You'd have a pad, or maybe a shallow tray, where, at
night, you put your small portable devices - phone, music player, PDA, electric razor - for recharging.
It would need to be standardized, so that all the gadgets could use a common system. You'd have one tray at home, probably in the bedroom. A travel version roll-up pad with a cord and wall transformer would be necessary. Deployment in business hotels would be useful.
Nice analysis. I've pointed out before that eBay prices on the PS3 were in a screaming dive within days of launch, but this uses enough data to really make that clear.
We went through this with the XBox 360, but with more speculators. People were trying to unload those things on eBay for months, finally at prices below retail.
The "secret reserve price" thing on eBay is a big part of the problem. That encourages overpriced items and wastes buyer time on auctions doomed to fail. Sellers like it, because they don't have to compete against each other on reserve price, but it probably reduces the number of successful transactions.
It's surprising how few sites have a decent rating system, like Slashdot. It certainly helps. (It would help even more if the editors were also rated, of course.)
I was just thinking yesterday of doing a paranoia web page, aggregating warnings from various sources.
US DHS terrorism threat level. ("Code Yellow, or Elevated." today.)
DoD InfoCon threat level. ("INFOCON level 4, "Increased Vigilance in Preparation for Operations or Exercises." today.)
California Office of Emergency Services warnings ("...FLASH FLOOD WATCH IN EFFECT FROM TUESDAY MORNING THROUGH LATE
TUESDAY NIGHT FOR THE NORTHERN AND CENTRAL PORTIONS OF THE SAN
FRANCISCO BAY AREA..." yesterday.)
The PS3 is indeed innovative. The Cell is the first non-shared memory multiprocessor in a mass market product. Whether this is a good idea or a dead end, like the Itanic, remains to be seen.
I personally think that the hassles of a non-shared memory machine outweigh the advantages, especially when the amount of memory per processor is on the low side. The XBox 360 is a 3-CPU shared memory multiprocessor, and presents no new programming problems. Historically, non-shared memory multiprocessors are very hard to program. The Ncube, the BBN Butterfly, and the Transputer all had that problem. However, enough PS3 machines are deployed that the effort is being expended to hammer through that problem.
It may not matter, though. The idea was supposed to be that the Cell processor would result in a cheaper machine than the competition's. That didn't work out.
Windows 2000 Service Pack 2 was the last Microsoft operating system where the user was in control. With SP3, Windows Update could make your machine a slave to the mothership in Redmond. In XP, most machines were slaved to Redmond. Remember the day Microsoft rebooted everybody by remote control via Windows Update? Vista just continues the trend - you will obey the commands from the mothership, or your machine stops working.
in quite a few cases, it might well be worth it- a multi-billion dollar merger, a head of state's emergency meeting, etc, etc.
Unlikely. That's what phones are for. Mergers aren't rush jobs, anyway; they take weeks to months to set up, half the time they fall through, and most of the time they lose money for the stockholders. There are some efforts underway to design a supersonic business jet, but the price has to be no more than 2x that of a comparably sized Grumman Gulfstream for it to sell. They're trying to get the operating cost down to business class fare levels, which, for a 14-seat plane, isn't bad.
Warren Buffet once went on an Alaska cruise, during which one of his companies had a crisis that kept him on a satellite phone for hours. He's interested in buying fifty of the supersonic bizjets for his NetJets rental operation. If they work.
In California, we have our governor, the Terminator, coming out against violent video games. Arnold does have his amusing moments.
But to really push the critic's buttons in the US, you have to have sex. No game publisher in the US would publish the stuff Illusion in Japan sells. "Battle Raper", "Sexy Beach", and "Artificial Girl".
Typical bugfix report: "Breast slider 1.5 download: With Ver1.0 was not possible, "it rubs", the chest and the nipple "it picks", "you play with the both hands", and so on colorful action was added. With adding these, expression conduct voice of the girl substantially power rise!
The skill which thinks the girl and the breast you are shy please do freely.
In addition, with Ver1.5 as been able to look at the girl to every nook and cranny, it reached the point where polygon of rear side of [bichimatsuto] [bichichiea] goes out."
Illusion apologizes for the delay in shipping the "Sexy Santa" module, which slipped to January 9, 2007. They're also having a contest - best sexy screenshot wins a microwave oven.
I'd love to see an operating IBM 602A again. That was the last and best of the commercial electromechanical punched card calculators. Plug board-wired, mechanical add, subtract, multiply, and divide, punched card input and output.
First shipped in 1948, and in commercial use well into the 1970s. 0.000003 MIPS.
A full tab shop was an IBM 82 sorter, a IBM 407 tabulator, a IBM 514 reproducer/summary punch, a IBM 77 collator, and a IBM 602A calculator. Plus some IBM 024 or IBM 026 keypunches, of course. With that set of gear, and a really big supply of blank punch cards and fanfold paper, you could do invoicing, general ledger, and payroll. The machinery was slow, but highly reliable.
What maybe wasn't there as a warning sign was the fact that AI research would be even slower. And that it would be so disjointed as to have half the CS guys in ivory towers busy postulating all sorts of maths theorems as fundamental conditions for an AI, while completely ignoring the neurologists, anthropologists, and even stage magicians piling up evidence that the brains just don't work that way.
This is something Microsoft got right. Bill Gates was unimpressed with traditional AI, and kept Microsoft Research mostly out of it. But he saw that Bayesian statistics actually worked, and Microsoft went heavily into that area. There's quite a bit of statistics-based AI in Microsoft products, from the grammar checker in Word (yes, it really is diagramming sentences) to the ordering of online help questions based on the likelihood of the answers. Today, most of the better work in AI is statistics-based, and the hard problems, like unstructured vision, are starting to yield to work in the field.
The combination of statistical techniques and sheer compute power works better than abstraction and mathematical logic on real world problems.
Stanford AI spent two decades on the wrong track. Not until they got a new generation of faculty did the place get unstuck. I used to call the second floor of the Gates Building "the place where AI went to die", and for a decade, it was.
There are too many websites, changing too rapidly, for volunteer raters to keep up. It's too labor-intensive. In a narrow area (like, say, hotels) it might work, and it's been done.
Look, when posting blogodreck about something somebody wrote, link to the actual article. This is supposedly about an article written by Prof. Nigel Smart at the University of Bristol. And it doesn't have a link to the article, or any useful reference to it. It doesn't even link to Prof. Smart's home page.
Here's Prof. Smart's home page.. He's a cryptographer, and one of the people behind elliptic curve cryptography, one of the alternatives to prime-number based systems. But I can't find any reference to risks of social networking on his pages.
So all we really have is some unknown blogger saying the obvious.
Advertising for prescription drugs used to be illegal. After that was "deregulated", it grew to twice the cost of drug R&D. There is now one pharmaceutical sales rep for every four doctors in the US.
Until the 1980s, drugs developed at Government expense went into the public domain immediately. Now, pharmaceutical companies can buy rights to government-developed drugs.
Big Pharma has negotiated several special deals to extend patent lifetimes. Patents are extended by the time the FDA spends evaluating the drug. And then there's a "proprietary rights in drug testing data" thing, which means that the company which did clinical testing gets an exclusion right against generic makers which can outlast the patent. And then there's a special extension of exclusivity deal if a drug company pushes an existing drug through clinical testing for children, which can extend the patent life.
But when the patent runs out, the price goes way down. Claritin used to be over $1/tablet; now the generic version is about $0.12 each.
The MIL-STD-810F test, procedure IV, calls for 29 drops of the test article while in its shipping box without functional damage. Think of that as soldiers unloading a truck in a hurry. Or baggage handling at some airports.
The operational tests are much milder. Procedure I, functional shock, is 40G for 11ms, 3x on each axis, with the unit running, without any operational glitches. Think of this as in use in an off-road vehicle bouncing over rough terrain, i.e. normal military usage.
Procedure VI, bench handling, is a 100mm drop test in normal orientation, power off, 4x. That's just dropping it on a table from 10cm.
3. The 3D UI [...] it's very tiring to keep your arms in the air while using a computer. Gestures do have their place, but not as the primary user interface for office systems.
But that's not the problem.
Autodesk tried this in the late 1980s, when it looked like "virtual reality" was the next big thing. It looked like a potential interface for 3D CAD. A system with multiple PC chassis and some special purpose hardware was built, and there was a VR test system,
the first on PC-type hardware.
The "hands in the air" thing isn't very precise. That's the problem. Much CAD work involves precise cursor positioning, and CAD tools are all about putting entities exactly where you want them, for which there are elaborate tools and modes. Various kinds of "snap" mechanisms help, but when drawing something complicated, getting close to the right entity to snap to can be tough. The 3D positional sensors from Polyhemus weren't accurate enough for that. (They never really were very good. I used to see them at graphics shows, usually with a dancer wearing the sensors and a screen showing the motion capture results. They'd never quite match, and the slow sensor rates were taking the life out of dance moves. Hard stops turned into mush. The sensors weren't even consistent; I'd ask the dancer to touch her fingers together, and on the screen, they'd be several inches from touching. That's hopeless for fine work. They didn't improve much over the years, either. The current generation of motion capture uses multiple cameras, and Polyhemus seems to have disappeared.)
The really wearing thing was VR goggles with lag. Turn head, wait for scene to settle, repeat. Everything needed to go about 10x faster for that to be tolerable.
The people of low-lying islands in Vanuatu, also in the Pacific, have been evacuated as a precaution.
Not all of Vanuatu, just some of the lower-lying islands. There are many Pacific islands just a few meters above sea level, most of them unpopulated. Some are only above sea level some of the time, like a tidal flat.
Roland the Plogger, cashing in on Xmas
on
The Physics of Santa
·
· Score: 0, Redundant
Roland the Plogger even tries to cash in on Xmas. How tacky.
Slashdot Mirror
The whole-house idea is overambitious, but, on a smaller scale, it would be a great way to recharge portable devices. You'd have a pad, or maybe a shallow tray, where, at night, you put your small portable devices - phone, music player, PDA, electric razor - for recharging. It would need to be standardized, so that all the gadgets could use a common system. You'd have one tray at home, probably in the bedroom. A travel version roll-up pad with a cord and wall transformer would be necessary. Deployment in business hotels would be useful.
Sun is going to have an impact on anything? Huh? Sun is imploding. Anybody want to buy their Fremont campus? It's empty.
What else is he expecting, a comeback of SGI?
Yes, looks bogus.
Domain names are so Web 1.0, anyway. In virtual reality, you have virtual real estate, like the "islands" of Second Life.
Nice analysis. I've pointed out before that eBay prices on the PS3 were in a screaming dive within days of launch, but this uses enough data to really make that clear.
We went through this with the XBox 360, but with more speculators. People were trying to unload those things on eBay for months, finally at prices below retail.
The "secret reserve price" thing on eBay is a big part of the problem. That encourages overpriced items and wastes buyer time on auctions doomed to fail. Sellers like it, because they don't have to compete against each other on reserve price, but it probably reduces the number of successful transactions.
It's surprising how few sites have a decent rating system, like Slashdot. It certainly helps. (It would help even more if the editors were also rated, of course.)
I was just thinking yesterday of doing a paranoia web page, aggregating warnings from various sources.
A web page with a good-looking version of this info, suitable for display on large screen displays, would be useful.
Once, Microsoft rebooted everybody who had auto reboot turned off but Windows Update turned on. That's when it became clear who was in control.
The PS3 is indeed innovative. The Cell is the first non-shared memory multiprocessor in a mass market product. Whether this is a good idea or a dead end, like the Itanic, remains to be seen.
I personally think that the hassles of a non-shared memory machine outweigh the advantages, especially when the amount of memory per processor is on the low side. The XBox 360 is a 3-CPU shared memory multiprocessor, and presents no new programming problems. Historically, non-shared memory multiprocessors are very hard to program. The Ncube, the BBN Butterfly, and the Transputer all had that problem. However, enough PS3 machines are deployed that the effort is being expended to hammer through that problem.
It may not matter, though. The idea was supposed to be that the Cell processor would result in a cheaper machine than the competition's. That didn't work out.
Windows 2000 Service Pack 2 was the last Microsoft operating system where the user was in control. With SP3, Windows Update could make your machine a slave to the mothership in Redmond. In XP, most machines were slaved to Redmond. Remember the day Microsoft rebooted everybody by remote control via Windows Update? Vista just continues the trend - you will obey the commands from the mothership, or your machine stops working.
Obey or Die - brought to you by Microsoft.
in quite a few cases, it might well be worth it- a multi-billion dollar merger, a head of state's emergency meeting, etc, etc.
Unlikely. That's what phones are for. Mergers aren't rush jobs, anyway; they take weeks to months to set up, half the time they fall through, and most of the time they lose money for the stockholders. There are some efforts underway to design a supersonic business jet, but the price has to be no more than 2x that of a comparably sized Grumman Gulfstream for it to sell. They're trying to get the operating cost down to business class fare levels, which, for a 14-seat plane, isn't bad.
Warren Buffet once went on an Alaska cruise, during which one of his companies had a crisis that kept him on a satellite phone for hours. He's interested in buying fifty of the supersonic bizjets for his NetJets rental operation. If they work.
This sounds like a training aid for torturers. Attorney General Gonzales ("Mr. Torture Memo") would love this.
Sony's own store at the Metreon in SF just power cycles all the demo machines every half hour, all at the same time.
In California, we have our governor, the Terminator, coming out against violent video games. Arnold does have his amusing moments.
But to really push the critic's buttons in the US, you have to have sex. No game publisher in the US would publish the stuff Illusion in Japan sells. "Battle Raper", "Sexy Beach", and "Artificial Girl".
Typical bugfix report: "Breast slider 1.5 download: With Ver1.0 was not possible, "it rubs", the chest and the nipple "it picks", "you play with the both hands", and so on colorful action was added. With adding these, expression conduct voice of the girl substantially power rise! The skill which thinks the girl and the breast you are shy please do freely. In addition, with Ver1.5 as been able to look at the girl to every nook and cranny, it reached the point where polygon of rear side of [bichimatsuto] [bichichiea] goes out."
Illusion apologizes for the delay in shipping the "Sexy Santa" module, which slipped to January 9, 2007. They're also having a contest - best sexy screenshot wins a microwave oven.
It's all so normal over there.
I'd love to see an operating IBM 602A again. That was the last and best of the commercial electromechanical punched card calculators. Plug board-wired, mechanical add, subtract, multiply, and divide, punched card input and output. First shipped in 1948, and in commercial use well into the 1970s. 0.000003 MIPS.
A full tab shop was an IBM 82 sorter, a IBM 407 tabulator, a IBM 514 reproducer/summary punch, a IBM 77 collator, and a IBM 602A calculator. Plus some IBM 024 or IBM 026 keypunches, of course. With that set of gear, and a really big supply of blank punch cards and fanfold paper, you could do invoicing, general ledger, and payroll. The machinery was slow, but highly reliable.
What maybe wasn't there as a warning sign was the fact that AI research would be even slower. And that it would be so disjointed as to have half the CS guys in ivory towers busy postulating all sorts of maths theorems as fundamental conditions for an AI, while completely ignoring the neurologists, anthropologists, and even stage magicians piling up evidence that the brains just don't work that way.
This is something Microsoft got right. Bill Gates was unimpressed with traditional AI, and kept Microsoft Research mostly out of it. But he saw that Bayesian statistics actually worked, and Microsoft went heavily into that area. There's quite a bit of statistics-based AI in Microsoft products, from the grammar checker in Word (yes, it really is diagramming sentences) to the ordering of online help questions based on the likelihood of the answers. Today, most of the better work in AI is statistics-based, and the hard problems, like unstructured vision, are starting to yield to work in the field. The combination of statistical techniques and sheer compute power works better than abstraction and mathematical logic on real world problems.
Stanford AI spent two decades on the wrong track. Not until they got a new generation of faculty did the place get unstuck. I used to call the second floor of the Gates Building "the place where AI went to die", and for a decade, it was.
Problems:
Look, when posting blogodreck about something somebody wrote, link to the actual article. This is supposedly about an article written by Prof. Nigel Smart at the University of Bristol. And it doesn't have a link to the article, or any useful reference to it. It doesn't even link to Prof. Smart's home page.
Here's Prof. Smart's home page.. He's a cryptographer, and one of the people behind elliptic curve cryptography, one of the alternatives to prime-number based systems. But I can't find any reference to risks of social networking on his pages.
So all we really have is some unknown blogger saying the obvious.
Advertising for prescription drugs used to be illegal. After that was "deregulated", it grew to twice the cost of drug R&D. There is now one pharmaceutical sales rep for every four doctors in the US.
Until the 1980s, drugs developed at Government expense went into the public domain immediately. Now, pharmaceutical companies can buy rights to government-developed drugs.
Big Pharma has negotiated several special deals to extend patent lifetimes. Patents are extended by the time the FDA spends evaluating the drug. And then there's a "proprietary rights in drug testing data" thing, which means that the company which did clinical testing gets an exclusion right against generic makers which can outlast the patent. And then there's a special extension of exclusivity deal if a drug company pushes an existing drug through clinical testing for children, which can extend the patent life.
But when the patent runs out, the price goes way down. Claritin used to be over $1/tablet; now the generic version is about $0.12 each.
A PS3 just went for $630 on eBay. Almost down to retail. We'll probably see the speculators trying to dump unsold stock starting tomorrow.
That adds up to $4,244. They want $15,000 for this gear, installed. What's wrong with this picture?
The MIL-STD-810F test, procedure IV, calls for 29 drops of the test article while in its shipping box without functional damage. Think of that as soldiers unloading a truck in a hurry. Or baggage handling at some airports.
The operational tests are much milder. Procedure I, functional shock, is 40G for 11ms, 3x on each axis, with the unit running, without any operational glitches. Think of this as in use in an off-road vehicle bouncing over rough terrain, i.e. normal military usage. Procedure VI, bench handling, is a 100mm drop test in normal orientation, power off, 4x. That's just dropping it on a table from 10cm.
"Push on Leonovo to offer something comparable to...." like the IBM/Lenovo ThinkPad T60p ?
That's not really "supported" for Linux. Read IBM/Leonovo's "Linux Certification - What does it mean? " page:
"Known problem areas and problems experienced in the past have been:
And this is a manufacturer writing about their own product., telling customers that, out of the box, it won't work right. That's not support.
But that's not the problem.
Autodesk tried this in the late 1980s, when it looked like "virtual reality" was the next big thing. It looked like a potential interface for 3D CAD. A system with multiple PC chassis and some special purpose hardware was built, and there was a VR test system, the first on PC-type hardware.
The "hands in the air" thing isn't very precise. That's the problem. Much CAD work involves precise cursor positioning, and CAD tools are all about putting entities exactly where you want them, for which there are elaborate tools and modes. Various kinds of "snap" mechanisms help, but when drawing something complicated, getting close to the right entity to snap to can be tough. The 3D positional sensors from Polyhemus weren't accurate enough for that. (They never really were very good. I used to see them at graphics shows, usually with a dancer wearing the sensors and a screen showing the motion capture results. They'd never quite match, and the slow sensor rates were taking the life out of dance moves. Hard stops turned into mush. The sensors weren't even consistent; I'd ask the dancer to touch her fingers together, and on the screen, they'd be several inches from touching. That's hopeless for fine work. They didn't improve much over the years, either. The current generation of motion capture uses multiple cameras, and Polyhemus seems to have disappeared.)
The really wearing thing was VR goggles with lag. Turn head, wait for scene to settle, repeat. Everything needed to go about 10x faster for that to be tolerable.
The people of low-lying islands in Vanuatu, also in the Pacific, have been evacuated as a precaution.
Not all of Vanuatu, just some of the lower-lying islands. There are many Pacific islands just a few meters above sea level, most of them unpopulated. Some are only above sea level some of the time, like a tidal flat.
Roland the Plogger even tries to cash in on Xmas. How tacky.